concepts and representation
everyday concepts cannot be defined in terms of necessary and sufficient conditions
a rule may not be *necessary* in the way that not all the instances contain the rule (chair has legs---> but a bean bag) a rule may not be *sufficient* in the way that it does not fully describe the category (dog---> cat)
definitional approach
we decide whether something is a member of a category by determining whether it meets the definition of the category
generalizing
when we say that a particular instance belongs in a certain category. Generalizing our conceptual knowledge to include that particular instance
limits of similarity based classification
-Exemplar models: assumes that similarity relations determine categorization, examples that are most similar to the item to be classified have the greatest influence on categorization -problems with categorization depending on similarity relations: similarity too variable & influenced by context similarity assessment and classification have been shown to be independent
context dependent nature of similarity
Meldin, goldstone and gentner- participants either saw A or B or B and C w/ the task of listing common and distinctive properties.Similarity depends on context
mismatch between similarity and categorization
Rips (1989)- participants asked to imagine a circular object with a 3in diameter. Categorization: asked whether this object was more likely to belong to the category of quarters or pizza Similarity: asked whether the object was more similar to the category of quarters or the category of pizzas *results*- categorization group judged the object more likely to belong in category of pizzas but more similar to category of quarters This demonstrates a dissociation between similarity judgement and category judgement, categorization cannot be reduced to similarity.
collins & quillian
a canary is a bird v. a canary is an animal The time is takes for a person to retrieve info about a concept is determined by the distance that must be traveled through the network. Statements that require further travel result in longer reaction times.
rosch mervis et al
common features experiment (global, basic, specific) the more specific the category label, the more features would be identified that are common all instances. Basic level is privileged.
Exemplar approach
concept is represented by multiple examples (rather than a single protoype) examples are actual members of the category to categorize compare the new item to stored examples representation is not abstract *a new exemplar is classified based on its similarity to stored representations of category instances* -takes into account atypical cases, explains typicality effect, easily deals w/ variable categories
categorizaton in infants
concepts are pre-semantic, conceptual development does not depend on language familiraztion/novelty preference procedure using eyetracking, infants look longer @ novelty objects when presented w/familiarized objects
hierarchical organization
concepts can be hierarchically organized larger more general categories are divided into smaller specific categories, creating a number of levels of categories -Rosch provided evidence for the idea that basic level is psychologically privileged
3 representational paradigm
definitonal approach prototype theory exemplar theory
protototype v. exemplar
exemplars work better for smaller categories prototype work better for larger categories
category
groups of objects that belong together b/c they belong to the same class of objects There are individual instances (exemplars) of a category
rosch's prototype experiment
high prototypicality- category member closely resembles category prototype ex. bird= category low prototypicality- category member odes not closely resemble category prototype bird=pernguin *typicality effect*- prototypical objects are processed preferentially. Highly prototypical objects are judged more rapidly (smith et al)
family resemblance
idea that family resemblance allow for some variation w/in a category. Suggest that categorization is based on how similar an object is to stored representation. Leads to the view that categorization is similarity based)
exemplar approach
involves determining whether an object is similar to all the stored exemplars in memory
prototype approach
membership in a category is determined by comparing the object to a prototype that represents the category
concept
mental representation
categorization
process by which things are placed into groups. Hidden mental processes (concept learning) v. observable responses (categorization)
concept learning
process of forming concepts
Prototype Approach
the typical member. Our representantion is composed of those feature values that are most frequently encountered in a category instance. Winner takes all determine which feature values are most frequently encountered the prototype does not have to be a member of the category
represena (tional paradigms
theories of concept learning aim to: *account for classification performance* (models must make accurate predictions on learning tasks) *suggest how categorical info might be organized in the mind*
Definitional Approach
there are defining features that act like criteria or rule for determining category membership -concepts are defined in terms of necessary and sufficient features *singly necessary- every instance of the concept must have the chosen defining property *jointly sufficient- every entity containing the defining properties must be an instance of the concept